Well Fracturing

Well Fracturing

Review

One of the most common methods of well stimulation is hydraulic fracturing.

This topic introduces the concepts, methodologies and considerations
required for a fracturing operation.

Content
Introduction

Fracturing is a well stimulation method in which large conductive fractures are created in
the formation around the wellbore, reaching far into the formation. Exerting pressure in
the wellbore that exceeds the fracture initiation pressure of the formation creates the
fractures. Fractures are initiated, propagated and held open as long as the injection
proceeds. When injection ceases, the fracture tends to close again. Two methods are
used to keep the fracture open after injection stops:
• propping the fracture using materials that are transported with the injected fluids;
• acid etching along the fracture wall – acid fracturing, which will not be discussed
in this topic.

It is generally understood that fracturing results in a single fracture orientated in the

plane of the wellbore. In shallow formations a horizontal fracture may occur.

Impact on Performance

When a well is fractured, the effect on well performance is equivalent to an enlargement

of the wellbore radius. For a constant pressure model, the rate behaviour that the
stimulation yields is an immediate increase in rate (as much as 10 or 100 times)
followed by a rapid rate decline. The early increase is often referred to as ‘flush
production’. Although this implies a short-lived production, the behaviour may last
several years in low permeability wells.
From a mathematical point of view, the impact of performance from hydraulics
fractures can be classified according to one of three models:
• infinite conductivity model – assuming no pressure loss in the fracture;
• uniform flux model – assumes a slight pressure gradient in the fracture;
• finite conductivity model – assumes constant and limited permeability in the
fracture from proppant crushing or poor proppant distribution.

Mathematical studies have shown, that an apparent wellbore radius can be used for the
infinite conductivity and uniform flux models, and in conjunction with the radial flow
model, these can adequately describe the production behaviour. Equations 1 and 2
below give the apparent wellbore radius for the infinite conductivity model and uniform
flux model respectively:
xf
Equation 1 rwa =
2

© IIT Madras 2011 1

 Well Fracturing

xf
Equation 2 rwa =
e
Where:
rwa apparent well radius
xf fracture width
e 2.718 (ie, ln(10))
For the finite conductivity model, the behaviour of the fracture is quantified by a
dimensionless fracture conductivity:
kf ⋅ W
Equation 3 FcD =
k ⋅ xf

Where:
FcD dimensionless fracture conductivity
kf fracture permeability
W fracture width (assume 0.01 ft average)
k formation permeability
xf vertical fracture half length

Layers of commercially available proppant materials can exhibit permeabilities from 120
to 2000 Darcies, and yet the fracture conductivity can be low and fracture length
calculated from test data using the infinite conductivity assumption is optimistic. Limited
fracture conductivity may result from:
• Incomplete proppant placement;
• Improper proppant size distribution;
• Improper proppant concentration in the fracture;
• Large closure stress resulting in proppant crushing;
• Fracture plugging by formation and fracturing fluid materials.

To help with a qualitative understanding of fracture conductivity, Table 1 below gives

characteristic values of the parameter detailed in Equation 3:
Table 1. Dimensionless Fracture Conductivity Parameters.
Quantity Value Characteristic
kf 10 D Poor
100 D Good
1000 D Excellent
kf.W 100 md-ft Poor
1000 md-ft Good
10000 md-ft Excellent
FcD <10 Poor
10-50 Good
>50 Excellent

Often the result of fracture treatment is poorer than that expected from the design. Test
data from the well will reveal this, but not the reasons for the poor result. As an example,
consider a fracture, which extends vertically across a sandstone formation and into
shale layers above and below. As a result of gravity, the proppant tends to settle at the
bottom of the fracture. When injection stops, the un-propped area begins to close as a

© IIT Madras 2011 2

 Well Fracturing

result of the horizontal stress due to the overburden. The result is an ineffective fracture
with limited extension, reduced area to flow and limited conductivity.
To study the effect of fracture conductivity and fracturing height on well performance, it
is possible to use results presented by McGuire and Skiora (1960) and Tinsley et al
(1969). Figure 1 shows the results of Tinsley’s study. The figure plots steady state flow
efficiency versus a fracture characterisation parameter for several ratios of fracture half
length to drainage radius.
Figure 1 Results from Tinsley et al (1969).

8
Flow Efficiency Ratio ln(re/rw)/ln(re/rwa)

0
0.1 1 10 100 1000 10000

X = 0.593 (hf/h)(xf/re)ln(re/rw)FcD

An important observation in Figure 1 is related to the relative importance of fracture

length and fracture conductivity. For a large X value (typical for low permeability
reservoirs) fracture length dominates the productivity gain and the fracture exhibits
infinite conductivity behaviour even if the conductivity has a finite value. At low values of
X, fracture conductivity (as well as fracture length) governs productivity gain.
The importance of this observation stems from the ability of the designer to trade-off
fracture conductivity for fracture length in designing and executing certain frac jobs.
In conclusion, any stimulated condition obtained by fracturing can be interpreted as a
radial well with an apparent wellbore radius. The results of a stimulation treatment can
be quantified as a skin factor:
r 
Equation 4 S = − ln wa 
 rw 
This however should be used cautiously – a post stimulation negative skin does not
imply that the stimulation will yield a prolonged increase in production rate. In fact, the
better the stimulation, the more rapidly the rate will decline with time. It is wrong then to
express the success of a stimulation in terms of a single constant flow efficiency
calculated from the initial post-stimulation production rate.

Geomechanics of Fracturing

A fluid injected into a formation modifies the pressure and stress distribution within the
formation, creating conditions that are favourable for the propagation of a fracture. In
addition, the injected fluid exchanges heat with the formation and also fluid will leak off
into the un-fractured reservoir. The injected fluid, because it contains proppant, is a

© IIT Madras 2011 3

 Well Fracturing

multiphase flow. The diagram in Figure 2 shows the interaction between these different
components.
Figure 2. Interaction of Hydraulics Fracture Modelling Components.

Fracture Mechanics

Fluid Mechanics Fluid Loss

Proppant Transport Heat Transfer

In this section consideration will be given to fracture propagation models and fluid flow
modelling.

Fracture Propagation Modelling

There are a number of fracture propagation modules including:

• two dimensional models;
• Perkins-Kern-Nordgren model;
• radial models;
• pseudo three dimensional model;
• three dimensional models.

The first model to simulate the propagation of fractures was developed by Khristianovic
and Zheltov in 1955. This two dimensional formulation is based on the assumption of
plane strains and two classes can be distinguished:
• plane strain condition in horizontal planes;
• plane strain condition in vertical planes.

By way of example, consider an infinite elastic medium and that each horizontal section
deforms independently from the others with no vertical strain. In this case, the horizontal
plane strain condition is in effect. All the z-components of the strain tensor, ε vanish.
Thus for the horizontal plane strain geometry, the fracture should deform independently
of the upper and lower layers. This would occur for free slippage in these layers, or
approximately represent a fracture with a horizontal penetration much smaller than the
vertical one. Such a geometry is shown in Figure 3 – it has a constant and uniform
height and a rectangular cross section.
A second simulation exists when there is a large confinement, hence the fracture is
limited to a given zone. Perkins and Kern (1961) and Nordgren (1972 considered the
plane strain assumption in vertical planes, so each vertical cross section deforms
independently of the others (PKN model). However the fracture width in vertical planes
are coupled through the fluid flow and continuity equations. Since there is no vertical
extension, the pressure is uniform and hence the shape of the fracture is elliptical. This
case approximates a fracture with a horizontal penetration much greater than the
vertical penetration.

© IIT Madras 2011 4

 Well Fracturing

Figure 3. Plane Strain Fracture Model.

Historically, the first type of hydraulic fracturing models considered a plane strain
solution in a horizontal plane relating to the fracture width and shape to the lateral
pressure distribution. Barenblatt (1962) described the notion of an equilibrium crack –
the fracture tip tends to zero at the crack tip. The width profiles associated with a given
pressure distribution are then given by the integral relation derived by England and
Green (1963)
Geertsma and de Klerk developed a tractable version of the model, considering the
size of the unpressurised zone near the crack tip. If σ 3 is the minimum horizontal stress
(the fracture will propagate in the direction perpendicular to it) and the excess pressure
is ∆p, distribution along the fracture is related to the fluid flow rate.
To solve the coupled problems of elasticity and fracture fluid flow, an iterative process
is required:
• assuming a pressure distribution in the fracture, determine the fracture width;
• given the fracture width distribution, compute the pressure from the fluid flow
equations;
• compare the assumed and computed pressures and if they differ, iterate.

To help simplify the fist step of the process, a step variation in the pressures along the
fracture can be assumed, thus:
Equation 4 p(x ) = p wi for 0 ≤ x ≤ x f (1 − ε L )

Equation 5 p(x ) = 0 for x f (1 − ε L ) ≤ x

Where :
xf vertical fracture half length
εL relative size of the zone not penetrated by the fluid
π σ 
Equation 6 εL = sin⋅  ⋅ 3 
 2 p wi 
Where:
Pwi wellbore injection pressure
σ3 minimum principal stress

© IIT Madras 2011 5

 Well Fracturing

These analytical relationships include a large number of assumptions:

• Constant injection rate;
• Constant fluid properties;
• No leak off.

To simulate the treatment conditions for which all these parameters may vary with time,
complex numerical solutions have been developed. The simulators are based on
different types of techniques including:
• Finite differences;
• Finite elements;
• Boundary integrals.

The selection of a particular method depends on different criteria including the

robustness and accuracy, ie, whether the technique will give correct results in most of
the simulations with a minimum number of external numerical controls.

PKN Model

The second type of model used to simulate the propagation of vertical hydraulic
fractures was presented by Perkins and Kern (1961 and later improved upon by
Nordgren (1972) who allowed for variations in the flow rate along the fracture. The
primary assumption of this model is that the fracture length is greater than the height.
Assuming that there is no flow in the vertical direction, the pressure in a vertical cross
section is constant, and the fracture has an elliptical shape.
The model proposed by Perkins and Kern gave a relationship between the width and
the excess pressure in a given cross section:
2∆PWi (x, t )
Equation 7 w (x, t ) = .h
E'
Where:
E’ plane strain modulus

Radial Models

The two previous models considered the propagation of a vertical fracture with a given
height. In some cases the vertical stress is lower that the horizontal stress causing the
fracture to propagate in a horizontal or inclined direction. By coupling the continuity,
elasticity and fluid flow equations, using an integral expression, the fracture width is
calculated from:
1
 µ  4
Equation 8 w = 2.19 ⋅  ⋅ qi ⋅ R 
 2 ⋅ E' 
Where:
R fracture radius

And the fracture volume is given by:

8⋅π
Equation 9 V= ⋅ w ⋅ R2
15

© IIT Madras 2011 6

 Well Fracturing

Pseudo Three Dimensional Models

To simulate vertical and lateral propagation of a vertical fracture, the assumption of a

constant and uniform height has to be removed. Therefore the problem needs to be
solved in three-dimensional space. This requires extensive computational times and
hence limits the practicality of the general use of such models for fracturing treatment
designs.
The basic concept of a pseudo three-dimensional model is the same as the PKN
model – that is, vertical planes deform independently. However, the height of the
fracture depends on the position along the fracture and the time. A vertical fracture will
grow in a layered medium as a function of the layer properties as well as the
characteristics of the injected fluid. The following factors affect growth:
• apart from linear variation in stress with depth due to the gradients, in-situ
stresses depend also on the lithology of the formation, tectonic components, and
pore pressure. The stress variations are generally the major factor for fracture
containment;
• elastic moduli – contrasts in Young’s modulus and Poisson’s ratio generally
impede the height growth but usually cannot stop a fracture;
• when the fracture height grows into a layer with a relatively high fluid loss, the
high rate loss will retard additional propagation through the zone.

Fracturing Fluids

The fracturing fluid is a critical component of the hydraulic fracture treatment. Its main
functions are to open the fracture and to transport the propping agent along the length of
the fracture. Consequently, the viscous properties of the fluid are often considered the
most important. However, successful hydraulic fracturing treatments require that the
fluids have some other special properties.
The first fracturing fluids were oil based and in the late 1950’s, water based fluids that
were thickened with Guar became popular. Concerns about damage to the formation led
to the utilisation of hydroxypropylguar based aqueous fluids in the 1970’s. Because of
their low cost, high performance and ease of handling, water-based fluids are the most
widely used fracturing fluids. Potential problems with water based fluids are:
• formation damage of water sensitive zones;
• pack damage caused by unbroken polymer and additives.

Fracturing Fluid Additives

A fracturing fluid is more often than not simply a liquid and viscosifying material such as
water and HPG polymer. Various additives are used to adjust pH, control bacteria,
improve high temperature stability, and break the fluid once the job is completed. Fluid
additives include:
• buffers;
• bactericides;
• stabilisers;
• breakers;
• surfactants;
• fluid loss additives.

Buffers are pH adjusting chemicals, which are added to aqueous fracturing fluids to
maintain a desired pH. The buffers, weak acids or bases or both are used in sufficient

© IIT Madras 2011 7

 Well Fracturing

quantities to maintain pH at a desired level even if an extraneous acid or base is

introduced through contaminated water or proppant.
Buffers are used to help with the hydration of polymers. For example, guar and HPG
are usually are usually treated to be dispersible and non-hydrating at high pH
The use of bactericides reduces the loss of viscosity due to bacterial degradation of
the polymer. Bacteria will not only ruin the gel by reducing the molecular weight of the
polymer, but also the can create sour wells. Once introduced into the well, some
bacteria can survive and reduce sulphate ions to H2S
Stabilisers are used to prevent degradation of polysaccharide gels at temperatures
above two hundred degrees Fahrenheit. The most common stabilisers are methanol and
sodium thiosulphate. There are certain health and safety issues when using Methanol
that must be considered. It usually makes up between five to ten percent of the fluid
volume. Sodium thiosulphate is more effective, and makes up ten to twenty percent of
the volume, however the mechanism of the action of these stabilisers is not fully
understood.
Thermal breaking of the polymer gel usually occurs in wells above 225 degrees
Fahrenheit. When treating lower temperature wells, a breaker should be added to the
fracturing fluid. Ideally, a gel breaker should be added to the fracturing fluid on surface
and should have no affect until pumping stops and the fracture starts to close.
A surface-active agent (or surfactant) is a material that at low concentrations adsorbs
at the interface between two immiscible substances, such as oil and water. Surfactants
are necessary to promote the stable formation of bubbles in foams and are used in poly-
emulsion fluids to stabilise the oil-in-water emulsion. Some bactericides and clay control
agents are also surfactants.
Good fluid loss control is essential for an efficient fracturing treatment. Several types
of materials are used to provide fluid loss control, but the effectiveness of the various
types depends on the type of fluid loss problem that is present:
• loss to the matrix;
• loss to microfractures;
• loss to macrofractures.

Fracture Proppant

Sand was the first material used as a proppant and since the late 1940’s several
materials have been used some of the unsuccessful efforts included aluminium pellets,
metal shot, glass beads, plastic beads and walnut shells. Some of the successful and
more commonly used propping agents today include sand, resin coated sand, ceramics
and high strength proppants (eg, sintered bauxite). Because of the large quantities and
wide variety of proppants, the American petroleum institute has established test
procedures for several proppant properties to distinguish the quality and usefulness of
each proppant (API RP56, 1983).

Sand

Premium sands come from Illinois, Minnesota and Wisconsin. These sands greatly
exceed API standards. They are commonly known as:
• Northern sand;
• White sand;
• Ottawa sand;
• Jordan sand;
• St. Peters sand;
• Wonewoc sand.

© IIT Madras 2011 8

 Well Fracturing

The specific gravity of sand is approximately 2.65.

Resin Coated Sand

Resin coatings may be applied to improve proppant strength. The resin coating on the
proppant is usually cured during the manufacturing process to form a non melting, inert
film. When the grains crush the resin coating helps encapsulate the crushed portions of
the grains and prevents them from migrating and plugging the flow channel. Resin
coated sands usually have a specific gravity of about 2.55.

Intermediate Strength Proppants

Intermediate strength proppants (ISP) are fused ceramic proppants that have a specific
gravity between 2.7 and 3.3. ISP’s are mainly used for closure stress ranges between
5,000 psi and 10,000 psi.

Intermediate Strength Proppants

Sintered bauxite and zirconium oxide are high strength propping agents with a specific
gravity of about 3.4 or higher. Because of their greater cost, they are generally limited to
wells with very high closure stresses.

Properties of Proppants

The physical properties of propping agents that have an impact on fracture conductivity
include the following:
• proppant strength;
• grain size;
• grain size distribution;
• quality (amount of fines and impurities);
• roundness and sphericity;
• proppant density.

Proppant Strength

It is common practice to use the difference between the initial fracture gradient
(minimum insitu stress) and bottom hole flowing pressure to calculate the maximum
effective stress on the proppant. The potential for maximum crushing can occur initially
in the production of a well because the fracture gradient is the maximum and decreases
with the reservoir pressure depletion. Usually the bottom hole flowing pressure is held
constant and at a low value in order to maximise the production rate.
Sand should be used for fracturing formations with closure stresses less than 6,000
psi. In the range of 5,000 to 10,000 psi ISP should be used. For closure stresses above
10,000 psi a high strength proppant is required.

Grain Size and Grain Size Distribution

Dirty formations, or those subject to significant fines migration are poor candidates for
large proppants the finds tend to invade the proppant pack, causing partial plugging and
a rapid reduction in permeability. In these cases, smaller proppants, which resist the
invasions of fines, are more suitable.

© IIT Madras 2011 9

 Well Fracturing

Quality

Grain size distribution and proppant quality are closely related. A high percentage of
smaller grains or impurities can have the same effect on the proppant pack permeability
as invading fines.

Roundness and Sphericity

Proppant grain roundness is a measure of the relative sharpness of grain corners, or of

grain curvature. Particle sphericity is a measure of how close the proppant particle or
grain approaches the shape of a sphere. Stresses on the proppant grains are more
evenly distributed when the grains are round and about the same size, resulting in
higher loads before grain failure occurs.

Proppant Density

Proppant density has an influence on proppant transport and placement. High density
proppants are more difficult to suspend in the fracturing fluid and to transport in the
fracture.

© IIT Madras 2011 10